Re: [PATCH] powerpc/booke64: wrap tlb lock and search in htw miss with FTR_SMT

Re: [PATCH] powerpc/booke64: wrap tlb lock and search in htw miss with FTR_SMT

[Scott Wood]

On Fri, 2014-05-30 at 17:59 +0300, Laurentiu Tudor wrote:
> Virtualized environments expose a e6500 dual-threaded core
> as two single-threaded e6500 cores.

s/expose/may expose/

I'll fix when applying.

>  Take advantage of this and get rid of the tlb lock and the trap-causing tlbsx in
> the htw miss handler by guarding with CPU_FTR_SMT, as it's
> already being done in the bolted tlb1 miss handler.

It's also possible that some non-virtualized workloads may also benefit
from getting rid of the lock more than they benefit from enabling SMT
(or perform better without SMT for other reasons).

-Scott

Patch not showing up in patchwork (was Re: [PATCH] powerpc/booke64: wrap tlb lock and search in htw miss with FTR_SMT)

[Scott Wood]

On Fri, 2014-05-30 at 17:59 +0300, Laurentiu Tudor wrote:
> Virtualized environments expose a e6500 dual-threaded core
> as two single-threaded e6500 cores. Take advantage of this
> and get rid of the tlb lock and the trap-causing tlbsx in
> the htw miss handler by guarding with CPU_FTR_SMT, as it's
> already being done in the bolted tlb1 miss handler.
> 
> As seen in the results below, measurements done with lmbench
> random memory access latency test running under Freescale's
> Embedded Hypervisor, there is a ~34% improvement.
> 
> Memory latencies in nanoseconds - smaller is better
>     (WARNING - may not be correct, check graphs)
> ----------------------------------------------------
> Host       Mhz   L1 $   L2 $    Main mem    Rand mem
> ---------  ---   ----   ----    --------    --------
> smt       1665 1.8020   13.2    83.0         1149.7
> nosmt     1665 1.8020   13.2    83.0          758.1
> 
> Signed-off-by: Laurentiu Tudor <Laurentiu.Tudor@freescale.com>
> Cc: Scott Wood <scottwood@freescale.com>
> ---
>  arch/powerpc/mm/tlb_low_64e.S | 4 ++++
>  1 file changed, 4 insertions(+)

Any idea why this patch isn't showing up on the mailing list?  I see my
reply in the archives, but not the original patch.

-Scott